Derrick J Brashear wrote: > >fine, so i'll guess this function calls one of our ops (getattr?) and >somehow loses, because EACCES shouldn't be coming from afs_lockctl (at >least not directly) > >Can you fstrace both cases?
Here's the sample C program I'm using: #include <sys/file.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> char *fname = "/s/db-apps/lib/GradApp/lib/GradApp.pbk"; main () { int fd, ret; fd = open (fname, O_RDONLY); ret = flock (fd, LOCK_SH); exit (0); } I think I have the helpful part of the fstrace. Tell me if you need more. Box 1 (the flock succeeds): <snip> time 119.118525, pid 29453: d_delete inode 0xe13ee698 d_name lib/GradApp time 119.118525, pid 29453: Access vp 0xe1077d3c mode 0x40 len 0x800 time 119.118525, pid 29453: Access vp 0xe13ef034 mode 0x40 len 0x800 time 119.118525, pid 29453: Access vp 0xe13f036c mode 0x100 len 0x23e8 time 119.118525, pid 29453: Open 0xe13f036c flags 0x0 time 119.118525, pid 29453: Open 0xe13f036c flags 0xf423f time 119.118525, pid 29453: Close 0xe13f036c flags 0x0 time 119.118525, pid 29453: d_delete inode 0xe13f036c d_name lib/GradApp.pbk AFS Trace Dump - Completed Box 2 (the flock fails): <snip> time 962.599815, pid 21833: d_delete inode 0x420bdf5c d_name lib/GradApp time 962.599815, pid 21833: Access vp 0x4204de4c mode 0x40 len 0x800 time 962.599815, pid 21833: Access vp 0x4203b168 mode 0x40 len 0x800 time 962.599815, pid 21833: Access vp 0x422de2c4 mode 0x100 len 0x23e8 time 962.599815, pid 21833: Open 0x422de2c4 flags 0x0 time 962.599815, pid 21833: Open 0x422de2c4 flags 0xf423f time 962.599815, pid 21833: Analyze RPC op 13 conn 0x2128b4c0 code 0xd user 0x417529fa time 962.599815, pid 21833: Returning code 13 from 1 time 962.599815, pid 21833: Returning code 13 from 3 time 962.599815, pid 21833: Close 0x422de2c4 flags 0x0 time 962.599815, pid 21833: d_delete inode 0x422de2c4 d_name lib/GradApp.pbk AFS Trace Dump - Completed Dave _______________________________________________ OpenAFS-info mailing list [EMAIL PROTECTED] https://lists.openafs.org/mailman/listinfo/openafs-info